Auto-sync from demo at Mon Dec 29 09:46:19 UTC 2025

app.py

@@ -160,6 +160,7 @@ def run_graphgen(params: WebuiParams, progress=gr.Progress()):
             "op_name": "generate",
             "type": "map_batch",
             "dependencies": ["partition"],
+            "save_output": True,
             "execution_params": {"replicas": 1, "batch_size": 128},
             "params": {
                 "method": params.mode,
@@ -168,14 +169,17 @@ def run_graphgen(params: WebuiParams, progress=gr.Progress()):
         }
     )
 
-    config = {"global_params": {"working_dir": working_dir}, "nodes": nodes}
+    config = {
+        "global_params": {
+            "working_dir": working_dir,
+            "graph_backend": "kuzu",
+            "kv_backend": "rocksdb",
+        },
+        "nodes": nodes,
+    }
 
     try:
         # 4. Initialize and Run Engine
-        # Initialize Ray if not already running (Engine handles this mostly, but good for safety)
-        if not ray.is_initialized():
-            ray.init(ignore_reinit_error=True, log_to_driver=True)
-
         engine = Engine(config, operators)
 
         # Start with an empty dataset to kick off the pipeline

graphgen/engine.py

@@ -7,15 +7,12 @@ from typing import Any, Callable, Dict, List, Set
 
 import ray
 import ray.data
-from dotenv import load_dotenv
 from ray.data import DataContext
 
 from graphgen.bases import Config, Node
 from graphgen.common import init_llm, init_storage
 from graphgen.utils import logger
 
-load_dotenv()
-
 
 class Engine:
     def __init__(

graphgen/run.py

@@ -6,6 +6,7 @@ from typing import Any, Dict
 
 import ray
 import yaml
+from dotenv import load_dotenv
 from ray.data.block import Block
 from ray.data.datasource.filename_provider import FilenameProvider
 
@@ -13,6 +14,8 @@ from graphgen.engine import Engine
 from graphgen.operators import operators
 from graphgen.utils import CURRENT_LOGGER_VAR, logger, set_logger
 
+load_dotenv()
+
 sys_path = os.path.abspath(os.path.dirname(__file__))
 
 

webui/app.py

@@ -160,6 +160,7 @@ def run_graphgen(params: WebuiParams, progress=gr.Progress()):
             "op_name": "generate",
             "type": "map_batch",
             "dependencies": ["partition"],
+            "save_output": True,
             "execution_params": {"replicas": 1, "batch_size": 128},
             "params": {
                 "method": params.mode,
@@ -168,14 +169,17 @@ def run_graphgen(params: WebuiParams, progress=gr.Progress()):
         }
     )
 
-    config = {"global_params": {"working_dir": working_dir}, "nodes": nodes}
+    config = {
+        "global_params": {
+            "working_dir": working_dir,
+            "graph_backend": "kuzu",
+            "kv_backend": "rocksdb",
+        },
+        "nodes": nodes,
+    }
 
     try:
         # 4. Initialize and Run Engine
-        # Initialize Ray if not already running (Engine handles this mostly, but good for safety)
-        if not ray.is_initialized():
-            ray.init(ignore_reinit_error=True, log_to_driver=True)
-
         engine = Engine(config, operators)
 
         # Start with an empty dataset to kick off the pipeline